Preliminary evidence suggests that 8-week training programs in meditation and exercise lead to reductions in incidence, duration and severity of acute respiratory infection (ARI) illness. In this parallel 3-group trial, women and men aged 30 to 69 will be randomized to: 1) an 8-week behavioral training program in mindfulness meditation, 2) an intensity, duration and location-matched 8-week exercise program, or 3) a wait-list observational control group. Recruitment will target those who do not exercise regularly, have not had training in meditation, and report levels of stress at or above average population levels. The primary outcome will be severity-weighted total days of ARI illness as assessed by self-reports on the validated Wisconsin Upper Respiratory Symptom Survey (WURSS-24). Weekly computer-assisted telephone monitoring for ARI illness will be operated by personnel blinded to intervention group. For each ARI illness episode, nasal samples will be tested with multiplex PCR (polymerase chain reaction) to identify viruses, and will be analyzed for neutrophil count, interleukin-6, and interleukin-8 concentrations. ARI-related visits to health care facilities and time lost from work and school will be documented, and assessed as secondary outcomes. Questionnaire measures assessing perceived stress, positive and negative emotion, self-efficacy, social support, sleep quality and general mental and physical health will be analyzed as potential mediators of causal pathways leading from behavioral training interventions to ARI illness outcomes. This will be a 5-year project, with 4 yearly cohorts of n=99 per cohort randomized into 3 groups of n=33 each. Assuming 9% loss to follow-up, the final sample size will be n=360 study participants, with n=120 in each comparison group. Enrollment, randomization and study interventions will begin in September. Participants will be monitored by weekly self-report through May. Summers will be used for data cleaning, preliminary analyses, and for recruiting the next year's cohort. Zero-inflated multivariate regression models will assess effects of meditation and exercise on ARI illness, health care utilization, and work absenteeism. Psychosocial measures will be assessed as potential mediators of effects of meditation and exercise on ARI illness. Generalized estimating equations, random-effects pattern-mixture models, and hierarchical linear models will be used to assess longitudinal effects, interactions, and covariate mediation.